**** BEGIN LOGGING AT Tue Feb 26 02:59:57 2019
Feb 26 08:16:30 <saidinesh5> @JaMa so should i be building thud branch?
Feb 26 08:16:47 <saidinesh5> Also i hope my 180GB disk space is enough for this...
Feb 26 08:27:49 <Herrie|Laptop> Morning!
Feb 26 08:28:04 <Herrie|Laptop> saidinesh5: For your tablet, do you use Android based build or direct linux with binary blobs?
Feb 26 08:28:16 <Herrie|Laptop> Because if direct linux with binary blobs, you don't need libhybris
Feb 26 08:28:27 <Herrie|Laptop> libhybris is used for talking to Android drivers
Feb 26 08:29:00 <saidinesh5> @Herrie|Laptop o/ i am trying to use the meta-intel layer directly... and i was hoping to find out where/what is pulling in that libhybris dependency (currently it says qtwebengine.. but not sure how)
Feb 26 08:29:47 <Herrie|Laptop> saidinesh5: Let me have a look
Feb 26 08:46:28 <saidinesh5> i did a grep for hybris in the raspberry pi layer and found nothing.. so wasn't sure if it was pulling it from shr or something
Feb 26 08:50:03 <Herrie|Laptop> saidinesh: I suspect it has to do with luneos-dev-package somehow
Feb 26 08:50:10 <Herrie|Laptop> Just I haven't found yet how exactly
Feb 26 08:57:40 <Herrie|Laptop> I think you might want to use luneos-dev-image
Feb 26 08:57:56 <Herrie|Laptop> That's what we do for the RPi devices:  $ MACHINE=raspberrypi3 bb luneos-dev-image
Feb 26 08:59:50 <Herrie|Laptop> Tofe: I think we might need to add this back? https://github.com/webOS-ports/meta-webos-ports/blob/pyro/meta-luneos/recipes-core/packagegroups/packagegroup-luneos-extended.bb#L133
Feb 26 09:21:10 <saidinesh5> Herrie|Laptop: wll try it out
Feb 26 09:30:00 <saidinesh5> WARNING: /src/LuneOS/webos-ports-env/webos-ports/meta-webos-ports/meta-luneos/recipes-webos/nyx-modules/nyx-modules.bb: Unable to get checksum for nyx-modules SRC_URI entry intel-corei7-64.cmake: file could not be found                                                                                                                                                                                                             | ETA:  0:00:35
Feb 26 09:30:00 <saidinesh5> WARNING: /src/LuneOS/webos-ports-env/webos-ports/meta-webos-ports/meta-luneos/recipes-webos/luna-sysmgr-conf/luna-sysmgr-conf.bb: Unable to get checksum for luna-sysmgr-conf SRC_URI entry luna-platform.conf: file could not be found  (need to fix these too..i think)
Feb 26 09:30:22 <saidinesh5> Herrie|Laptop: ERROR: Required build target 'luneos-dev-image' has no buildable providers. Missing or unbuildable dependency chain was: ['luneos-dev-image', 'packagegroup-luneos-extended', 'qtwayland-plugins', 'libhybris']
Feb 26 09:30:50 <saidinesh5> i think i have to manually  add some "provides" to my meta-intel-luneos i think
Feb 26 09:30:59 <saidinesh5> maybe an xserver or wayland (preferably the latter)
Feb 26 09:31:47 <saidinesh5> so once i am off work, will add stuff to that meta-intel-luneos i am making , that includes some file from meta-intel and adds more provides/removes
Feb 26 09:42:52 <Tofe> Herrie|Laptop: oh, right, of course
Feb 26 09:44:53 <Tofe> saidinesh5: could it be your intel tablet has an IGP directly supported by Mesa ?
Feb 26 09:45:49 <saidinesh5> Tofe: I think so.. that's why i added teh meta-intel layer
Feb 26 09:46:02 <saidinesh5> so need to figure out what's pulling in the libhybris and remove it
Feb 26 09:49:24 <Tofe> libhybris should only be pulled if you specify that virtual/EGL is provided by libhybris in the machine conf file
Feb 26 09:49:47 <saidinesh5> Ahhh
Feb 26 09:51:00 <Tofe> There are other possibilities (like a mistake in luneos-dev-package definition), but this one is the main one
Feb 26 09:51:11 <saidinesh5> https://git.yoctoproject.org/cgit.cgi/meta-intel/tree/conf/machine/intel-corei7-64.conf
Feb 26 09:51:15 <saidinesh5> this is my machine conf
Feb 26 09:57:32 <Tofe> Seems ok; you'll probably also need to add some things similar to qemu: https://github.com/webOS-ports/meta-webos-ports/blob/pyro/meta-luneos/recipes-core/packagegroups/packagegroup-luneos-extended.bb#L154
Feb 26 09:58:49 <Tofe> Just to be sure, did you already modify packagegroup-luneos-extended.bb ? In which case, what did you add/change ?
Feb 26 10:03:30 <JaMa> saidinesh5: for stable and testing builds we use pyro
Feb 26 10:03:35 <saidinesh5> i haven't modified anything except the BBLayers so far though
Feb 26 10:03:41 <saidinesh5> just added the meta-intel to the bblayers
Feb 26 10:03:49 <JaMa> rocko,sumo,thud,master aren't ready for end-users
Feb 26 10:03:58 <saidinesh5> Ahh I see
Feb 26 10:04:16 <JaMa> so it depends if you want to help with migration to newer Yocto or just get your intel support working in current pyro
Feb 26 10:06:33 <saidinesh5> well once I have a working build, I can help with migration I think
Feb 26 10:06:42 <saidinesh5> otherwise there will be too many unknowns right?
Feb 26 10:10:49 <Tofe> right
Feb 26 11:13:09 <Herrie|Laptop> Tofe: Should I PR the change?
Feb 26 12:17:12 <Tofe> Herrie|Laptop: it would be good to verify the build first, as it would basically mean revert that revert: https://github.com/webOS-ports/meta-webos-ports/commit/2e37a061d1ebb5772bcfbf28e9e1b2f329c31fe2
Feb 26 12:18:45 <Tofe> and to me it kind of indicates we might be forgetting something here :p
Feb 26 13:00:05 <Herrie|Laptop> Tofe: Well you said you fixed the modules, right?
Feb 26 13:00:17 <Herrie|Laptop> Or maybe they need some more fixing
Feb 26 13:00:29 <Herrie|Laptop> JaMa's reversal seems to be sumo/thud related?
Feb 26 13:23:53 <Herrie|Laptop> Looks somebody @ home was playing with the power button of the PC :P
Feb 26 13:24:01 <Herrie|Laptop> Need to disconnect it I guess ;)
Feb 26 14:04:44 <JaMa> Herrie|Laptop: reversal?
Feb 26 14:05:21 <Tofe> JaMa: s/reversal/revert/
Feb 26 14:06:06 <Tofe> Herrie|Laptop: yes, it should work, I just want to double check before doing one more ping-pong revert
Feb 26 14:07:13 <Herrie|Laptop> Tofe: Will try a build once builder is back online and let you know
Feb 26 14:07:16 <JaMa> I've forgot all about this commit to be honest
Feb 26 14:23:59 <Herrie|Laptop> JaMa: Stuff might or might not have been fixed with https://github.com/webOS-ports/nyx-modules-hybris/commits/master
Feb 26 18:16:15 <saidinesh5> anyone here?
Feb 26 18:16:59 <saidinesh5> so meta-luneui > recipies-qt > qt5 >  qtwayland_git.bbappend has a line like this:
Feb 26 18:17:00 <saidinesh5> EXTRA_PACKAGECONFIG ?= "libhybris-egl-server"
Feb 26 18:17:00 <saidinesh5> EXTRA_PACKAGECONFIG_qemuall = "drm-egl-server"
Feb 26 18:17:00 <saidinesh5> EXTRA_PACKAGECONFIG_rpi = ""
Feb 26 18:17:27 <saidinesh5> can i change add another EXTRA_intel-corei7-64="drm-egl-server" ?
Feb 26 18:19:28 <saidinesh5> ookay that seems to work
Feb 26 18:19:58 <Tofe> yes, you can :)
Feb 26 18:20:50 <Tofe> actually it's not great that it's located in meta-webos-ports
Feb 26 18:21:17 <saidinesh5> yeah.. but it is okay
Feb 26 18:21:26 <Tofe> that'll do for now
Feb 26 18:21:37 <saidinesh5> so i will have to make sure i am not building too much additional stuff
Feb 26 18:21:54 <saidinesh5> Tofe: is there a way to quickly get an overview of the changes i made?
Feb 26 18:21:59 <saidinesh5> like repo status or some such thing?
Feb 26 18:22:38 <Tofe> not repo, but git status in meta-webos-ports and meta-smartphone should be quite all you've done
Feb 26 18:22:52 <saidinesh5> i also edited conf/bblayers etc.. too
Feb 26 18:23:14 <saidinesh5> https://bpaste.net/show/0906646f9942
Feb 26 18:23:22 <Tofe> unfortunately it's not really centralized
Feb 26 18:23:47 <saidinesh5> Ahhh
Feb 26 18:24:06 <Tofe> git diff might also give details. But not for the files that were not added in git
Feb 26 18:24:23 <saidinesh5> would be nice if all these meta-* are added as git submodules or something and that way everything gets updated with git pull
Feb 26 18:24:27 <saidinesh5> but i digress
Feb 26 18:24:28 <saidinesh5> 1 sec
Feb 26 18:24:44 <saidinesh5> So do i need to have a meta-intel-luneos?
Feb 26 18:24:53 <saidinesh5> right now i don't think i have any stuff in there.. yet
Feb 26 18:25:25 <Tofe> I don't know if you need it, I just read earlier that you used intel's layer -- maybe I misunderstood
Feb 26 18:25:30 <saidinesh5> added:  to bblayers.conf ${TOPDIR}/meta-intel \
Feb 26 18:25:43 <saidinesh5> yeah i used the meta-intel , like there is a meta-raspberrypi
Feb 26 18:25:51 <Tofe> ok yes
Feb 26 18:26:03 <saidinesh5> but there is also a meta-rpi-luneos, seems to have luneos related overrides
Feb 26 18:26:21 <saidinesh5> so blindly created meta-intel-luneos .. but there is nothing in there yet
Feb 26 18:26:54 <saidinesh5> my machine=intel-corei7-64 directly is defined in meta-intel itself..
Feb 26 18:27:04 <Tofe> It could also fit in meta-smartphones, I guess. I'm never quite sure in which order our layers are used
Feb 26 18:27:12 <Tofe> JaMa: ^
Feb 26 18:27:47 <Tofe> You'll very probably need to refine your machine.conf for luneos
Feb 26 18:28:07 <saidinesh5> Ahh so there has to be a meta-intel-luneos then anyway right?
Feb 26 18:28:31 <Tofe> I never had to make one myself, but I guess it would make sense
Feb 26 18:29:35 <Tofe> you'll probably end up with similar things as in meta-rpi-luneos repo
Feb 26 18:29:57 <saidinesh5> I see
Feb 26 18:30:14 <saidinesh5> btw. how come you guys updated only qt5 to thud, and everything else is pyro? in  layers.txt
Feb 26 18:31:34 <Tofe> Though most of what's in there is to fix build issues, due to the weird locations of the egl headers
Feb 26 18:32:08 <saidinesh5> the meta-rpi-luneos right?
Feb 26 18:32:11 <saidinesh5> yep seems like it
Feb 26 18:32:15 <Tofe> yes
Feb 26 18:32:48 <Tofe> saidinesh5: meta-qt5 lives quite its own life; for pyro we went up to 5.11.3, as pyro is - so far - our daily working codebase
Feb 26 18:33:47 <Tofe> rocko/sumo/thud are more WIP, and therefore we try to couple them with the latest meta-qt5, even if there will be fixes to be done
Feb 26 18:33:57 <saidinesh5> ahh .. but this layers.txt has everything else on pyro and just Qt on thud
Feb 26 18:34:04 <saidinesh5> testing one
Feb 26 18:34:52 <Tofe> relying on pyro or thud doesn't necessarily change our luneos-related recipes: in a perfect world, it would work as-is on both.
Feb 26 18:35:33 <saidinesh5> Ahh but the rest of the packages from pyro - like say network manager qt/what webos uses there may depend on a specific qt right?
Feb 26 18:35:50 <Tofe> same for meta-qt5: ideally we could use any recent version of it (Qt 5.9, 5.10, 5.11, 5.12...) without any change on our side. Of course, it's not so simple, but we try to minimize the work
Feb 26 18:36:00 <saidinesh5> Ahh..
Feb 26 18:36:19 <saidinesh5> So this layers.txt is okay? https://bpaste.net/show/f3723aeb4958
Feb 26 18:36:54 <Tofe> LGTM
Feb 26 18:38:22 <saidinesh5> (Y)
Feb 26 18:39:28 <saidinesh5> started bitbake: https://bpaste.net/show/0411ca416147
Feb 26 18:40:51 <Tofe> :)
Feb 26 18:41:30 <saidinesh5> i think i will have to buy more ram for this ryzen 7.. just 16gigs right now.. so added a swapfile of another 16
Feb 26 18:45:06 <Tofe> I also have 16Gb... but in my case the CPU is the bottleneck (i5 6600)
Feb 26 18:45:48 <saidinesh5> Ah.. for me with 16GB ram.. a chrome + firefox open.. building something like qtwebengine is *verrrry* risky
Feb 26 18:48:00 <saidinesh5> basically my old 2013 nexus 7is slowly dying.. so i need a decent touch optimized OS, which fits nicely in 2GB ram for this intel atom tablet.. I could go back to android... buuut an OS with gestures seems nicer
Feb 26 18:50:47 <saidinesh5> plus it would be ****verrrry*** nice to have something like this enabled in the browser: https://01.org/blogs/joone/2018/using-chrome-os-graphics-stack-intel-based-linux-desktops
Feb 26 18:51:16 <saidinesh5> on windows 10 that came with tablet, Edge performed quite nicer than firefox/chromium..
Feb 26 18:51:42 <saidinesh5> so with some patches like this, it would be quite nice to have this on linux too
Feb 26 18:53:55 <saidinesh5> not exactly sure how to make QtWebEngine take advantage of this though
Feb 26 18:55:01 <Tofe> well if the EGL drivers are optimized, it should benefit QtWebEngine
Feb 26 18:55:54 <Tofe> we don't have X, that's for sure ! :)
Feb 26 18:56:37 <saidinesh5> oh this machine pulls in X no?
Feb 26 18:56:40 <saidinesh5> *MACHINE
Feb 26 18:57:03 <saidinesh5> qemu image too uses X right?
Feb 26 18:57:42 <saidinesh5> https://git.yoctoproject.org/cgit.cgi/meta-intel/tree/conf/machine/intel-corei7-64.conf
Feb 26 19:00:39 <saidinesh5> So i have to add this stuff to my meta-intel-luneos ? https://bpaste.net/show/d1b6c2b90ee2
Feb 26 19:11:24 <Tofe> nope, our compositor is wayland-only, we don't need X
Feb 26 19:12:05 <Tofe> defining XSERVER won't necessarily make it used/installed; though I quite don't know the details...
Feb 26 19:12:32 <Tofe> We just need EGL drivers, basically.
Feb 26 19:15:06 <Tofe> ( ... with the "EGL_WL_bind_wayland_display" extension )
Feb 26 19:17:43 <saidinesh5> Ahh I see
Feb 26 19:34:51 <JaMa> Tofe: I should limit which MACHINEs I'm building for unstable (rocko, sumo, thud) to make the sstate a bit smaller, should I do just tissot and qemux86 for now? or are other devices interesting for you as well right now (will enable more once we get some first ones working a bit)
Feb 26 19:35:30 <JaMa> the builder is still doing thud for last 2-3 days
Feb 26 19:36:08 <Tofe> for now that's fine; I'd be interested in hammerhead too, but I understand it doesn't share much with the rest, and I can very well built it myself too
Feb 26 19:37:13 <Tofe> JaMa: but anyway I rarely use your sstate, most of the time I'm better off rebuilding it locally than downloading it ;)
Feb 26 19:38:25 <saidinesh5> I can use your sstate? O_O
Feb 26 19:38:49 <saidinesh5> or wait.. i dont think that would be possilbe when machines are different right?
Feb 26 19:40:17 <Tofe> saidinesh5: well, you're using x86-64, so maybe there's a possibility of reusing qemux86-64's sstate
Feb 26 19:41:36 <Tofe> JaMa: with https://bpaste.net/show/0411ca416147 , saidinesh5 shares some common sstate with qemux86-64, right ?
Feb 26 19:42:19 <Tofe> though you just said you'll first only build qemux86 :p
Feb 26 19:45:35 <saidinesh5> Bah... Ran out of ram
Feb 26 19:45:40 <saidinesh5> System froze up
Feb 26 19:47:28 <Tofe> oh. I never get that, maybe you can change the build configuration ?
Feb 26 19:47:44 <saidinesh5> Yeah.. Waiting for the system to unfreeze
Feb 26 19:47:54 <saidinesh5> Trying to ssh from phone and kill Firefox
Feb 26 19:48:02 <Tofe> I have 4 thread + make j4
Feb 26 19:49:05 <Tofe> conf/local.conf: I have PARALLEL_MAKE = "-j 4" and BB_NUMBER_THREADS = "4"
Feb 26 19:49:14 <saidinesh5> Ah
Feb 26 19:50:01 <Tofe> My CPU is mostly 100% the whole time, so I think it's a pretty good config for me
Feb 26 19:50:37 <saidinesh5> Yeah.. will definitely have to get more ram for this machine.. otherwise it'll be very underutilized
Feb 26 19:53:28 <saidinesh5> Btw. I think sharing sstate with qemu x86_64 may not be possible... Because of the TUNE_FEATURES
Feb 26 19:53:42 <Tofe> oh... right...
Feb 26 19:53:42 <saidinesh5> Maybe noarch packages...
Feb 26 19:53:57 <Tofe> well that's not the most interesting part :p
Feb 26 19:54:02 <saidinesh5> Yup
Feb 26 19:55:39 <Tofe> Which Rizen do you have ?
Feb 26 19:55:49 <saidinesh5> 1700..
Feb 26 19:56:10 <saidinesh5> Ryzen 7 1700. got it in a clearence sale this december lol. so nice 8 cores 16 threads
Feb 26 19:56:38 <Tofe> oh, 16 threads, I see... You use -j 16 ?
Feb 26 19:56:53 <saidinesh5> bit bake used 16 jobs.. plus each job seems to be parallellized lol
Feb 26 19:57:40 <Tofe> I'd say try with 4 and -j8, it should already take most of it without topping the memory too much
Feb 26 19:57:56 <saidinesh5> yeah.. fortunately the swapfile i added saved the day!
Feb 26 19:57:58 <saidinesh5> system unfroze
Feb 26 19:58:24 <saidinesh5> 4908 tasks done out of 7185
Feb 26 19:58:48 <saidinesh5> i will let it finish.. it is not as if i sit around recompiling the same packages until we update from pyro right?
Feb 26 20:00:45 <Tofe> oh it's a 65W TDP ? Same as mine; I mainly selected it to consume not too much while being decent enough, and at the time AMD was doing quite poorly on that former side
Feb 26 20:01:43 <saidinesh5> my main powerhog is that RX580
Feb 26 20:01:59 <saidinesh5> that one gobbles up like 400W when fully running i think
Feb 26 20:02:19 <Tofe> I have a Nvidia 1050Ti, 75W, for the same reason :)
Feb 26 20:02:29 <saidinesh5> but for the first time ever .. 0 issues with graphics on linux with foss drivers lol
Feb 26 20:03:11 <saidinesh5> wow .. 75W sounds pretty sweet for a GPU
Feb 26 20:03:24 * Tofe doublechecks...
Feb 26 20:03:59 <Tofe> yup, that's the right number.
Feb 26 20:04:26 <saidinesh5> Oh wait ... Mine is 225W
Feb 26 20:04:35 <Tofe> *only* :)
Feb 26 20:04:38 <saidinesh5> I confused this with the Vega 56
Feb 26 20:04:41 <saidinesh5> Lol
Feb 26 20:05:25 <saidinesh5> My choices then were buy an rx580 here.. or make a friend from the US bring me one of those vega 56 for the same price I'd have paid here
Feb 26 20:05:35 <saidinesh5> For my rx580
Feb 26 20:06:03 <Tofe> But then at 400W the price of electricity enters the equation ;)
Feb 26 20:06:12 <saidinesh5> Yep
Feb 26 20:06:26 <saidinesh5> That's why decided against it
Feb 26 20:07:28 <saidinesh5> I mean the games i played with... I was literally playing 10 year old games on this Intel atom tablet .. And some games on 720p on that laptop gpu lol.. it's not as if I'll care about the graphics when i was used to it
Feb 26 20:09:49 <saidinesh5> so far this 580 played Hitman 2 very nicely .. so i am very happy... and it's opensource amd drivers => 0 headaches meddling with bumblebee etc.. breath of fresh air
Feb 26 20:09:53 <Tofe> :) I understand, until recently I just needed my Intel HD 530 on the i5, and it was mostly alright
Feb 26 20:10:11 <saidinesh5> HD 530 ? which game did you play?
Feb 26 20:11:26 <saidinesh5> or rrather which game made you upgrade to the 1050ti?
Feb 26 20:11:31 <Tofe> Civilization V, Fallout 3, Morrowind, Shank
Feb 26 20:11:34 <saidinesh5> Ahh
Feb 26 20:12:07 <Tofe> Now I can play Tomb Raider, Batman Arkham, things like that
Feb 26 20:12:18 <saidinesh5> yup.. Tomb Raider was surprisingly nice on linux
Feb 26 20:12:31 <saidinesh5> waiting for feral to release the linux version of the latest.. so i can buy it
Feb 26 20:12:39 <Tofe> :)
Feb 26 20:12:50 <saidinesh5> bought arkham during the last steam sale.. yet to play that .. spending a lot of time on this simulator
Feb 26 20:13:18 <saidinesh5> gotta admit.. really thankful to valve for literally changing the linux graphics/gaming scene
Feb 26 20:13:28 <Tofe> it plays very nice with wine -- I also bought it with that sale
Feb 26 20:13:38 <saidinesh5> Shadow?
Feb 26 20:13:47 <Tofe> I just played Origin so far
Feb 26 20:13:53 <saidinesh5> Ahh..
Feb 26 20:14:07 <Tofe> but I guess (hope?) that Shadow will play as nicely
Feb 26 20:14:18 <Tofe> with DXVK it's surprisingly fast
Feb 26 20:14:25 <saidinesh5> Shadow of the tomb raider.. i am waiting for feral to release  so my money directly goes to them..  the last two games.. i bought in sale and then feral released linux versions .. so i got them for free
Feb 26 20:14:37 <saidinesh5> yup. i have heard that too
Feb 26 20:15:17 <saidinesh5> hitman 2.. i played on windows.. totally loved hitman 1.. so didn't wait for feral this time.. i hear even that runs very well on wine/Steamplay on linux
Feb 26 20:15:19 <Tofe> Let's just wait for a Vulkan version of Chromium, eh :)
Feb 26 20:15:29 <saidinesh5> haha... is that even in their roadmap?
Feb 26 20:15:34 <Tofe> dunno
Feb 26 20:16:12 <Tofe> I don't even know if the latest Qualcomm CPUs provide Vulkan drivers
Feb 26 20:16:21 <saidinesh5> they do right?
Feb 26 20:16:41 <Tofe> I don't have a clue -- but I do hope so
Feb 26 20:16:42 <saidinesh5> i mean vulkan has been part of android since android 7 i think.. so any qualcomm chipset released since then has vulkan suppor
Feb 26 20:17:03 <Tofe> oh ? ok, I learned something tonight, good :)
Feb 26 20:17:35 <Tofe> But libhybris... I didn't see anything on that front
Feb 26 20:17:45 <saidinesh5> Ahh. true..
Feb 26 20:18:23 <Tofe> well, I guess it's not our top-priority either :D
Feb 26 20:18:39 <saidinesh5> but what userspace program on hybris side of OSes even use vulkan?
Feb 26 20:20:00 <saidinesh5> plus if it is an SDL program .. we can directly link it wiht bionic and it should work right?
Feb 26 20:21:08 <saidinesh5> is our QtWebEngine actually hardware accelerated?
Feb 26 20:24:25 <saidinesh5> on libhybris
Feb 26 20:25:02 <JaMa> saidinesh5: Tofe: with default config it should share a bit (mostly just native stuff and some low level deps, but if there is bbappend for qtbase, then not the rests (like qtwebengine and all big stuff)
Feb 26 20:25:53 <JaMa> the TUNE_FEATURES are the same "m64 corei7"
Feb 26 20:25:55 <saidinesh5> Ah .. I see.. how can it share though.. like TUNE_FEATURES are different right?
Feb 26 20:25:56 <saidinesh5> oh
Feb 26 20:25:59 <saidinesh5> i see
Feb 26 20:26:05 <Tofe> saidinesh5: QtWebEngine uses our GLESv2 drivers yes
Feb 26 20:26:13 <JaMa> ah no, I cannot read, sorry qemux86-64 is "m64 core2"
Feb 26 20:26:18 <saidinesh5> oh that's very nice..
Feb 26 20:26:20 <Tofe> As does QML also
Feb 26 20:26:26 <saidinesh5> yeah QML has to
Feb 26 20:26:32 <saidinesh5> just not sure about webengine though
Feb 26 20:26:57 <saidinesh5> so that's kind of why i totally ruled out sailfish os/nemomobile for this tablet
Feb 26 20:27:28 <saidinesh5> the browser.. the one core piece of software is totally unoptimized for it.. and when i tried to compile chromium/qtwebengine.. it wasjust a nightmare
Feb 26 20:27:41 <saidinesh5> old gcc.. half the packages qtwebengine needs aren't available in mer repos
Feb 26 20:28:08 <Tofe> SFOS's browser is currently a shame, I hope they find the time to upgrade it and/or to provide a recent QtWebEngine
Feb 26 20:28:11 <saidinesh5> so thought.. no point in redoing the packaging work.. someone else has done so many times before me.. might as well pick a distro that lets me skip that
Feb 26 20:28:52 <saidinesh5> yep.. wrote a little browser-let.. to mimic android's browser on sailfish os.. but even that's broken as of now
Feb 26 20:29:15 <saidinesh5> need to update my patch to lipstick for that
Feb 26 20:29:54 <saidinesh5> Tofe: btw. how's anbox looking on luneos?
Feb 26 20:32:31 <saidinesh5> i think with the MiA1 you guys should have pretty recent kernel.. so shouldn't be too hard to get it up and running
Feb 26 20:33:53 <Tofe> saidinesh5: status unknown so far, I'm more trying to fix the vkb issue than anbox right now
Feb 26 20:34:00 <saidinesh5> woot: 10: gnu-efi-3.0.5-r0 do_package - 20s (pid 3282)
Feb 26 20:34:00 <saidinesh5> 11: diffutils-3.5-r0 do_package_write_ipk - 5s (pid 4998)
Feb 26 20:34:00 <saidinesh5> 12: grub-efi-2.00-r3 do_configure - 5s (pid 6013)
Feb 26 20:34:00 <saidinesh5> 13: syslinux-6.03-r0 do_configure - 3s (pid 6571)
Feb 26 20:34:05 <saidinesh5> oh .. what vkb issue?
Feb 26 20:35:14 <Tofe> it looks like a touchscreen issue, either on kernel side or on userland side: somehow, we register two touches, so the text field gets focus and immediately after loses it, so the vkb just pops up and disappear
Feb 26 20:36:00 <saidinesh5> oh interesting...
Feb 26 20:36:04 <Tofe> at each boot, it either works well or not at all. It seems to be related to the way we initialize it -- maybe also related to the way we initialize hwcomposer
Feb 26 20:36:46 <saidinesh5> does hwcomposer come into picture for input?
Feb 26 20:36:58 <saidinesh5> i thought input happens via. /dev/input/*
Feb 26 20:37:07 <Tofe> only when we disable it after a timeout: it also deactivates the touchscreen
Feb 26 20:37:21 <saidinesh5> oh.. interesting
Feb 26 20:38:16 <saidinesh5> it = hwcomposer?
Feb 26 20:38:23 <Tofe> I've spent quite some hours so far, and with no precise leads to chew; it's a bit frustrating
Feb 26 20:38:24 <Tofe> yes
Feb 26 20:39:42 <Tofe> I've also tried to debug it on rosy (Mi 5), where it happens almost all the time, but without much success
Feb 26 20:40:04 <saidinesh5> interesting.. Mi 5 has a very well working sailfish port
Feb 26 20:40:13 <saidinesh5> so shouldn't be kernel right?
Feb 26 20:40:34 <Tofe> shouldn't be; also recovery is working very well, using the same kernel
Feb 26 20:40:42 <saidinesh5> yep
Feb 26 20:40:45 <saidinesh5> could it be just Qt?
Feb 26 20:41:00 <saidinesh5> like can you disable the compositor and run an SDL program or osmethign and check?
Feb 26 20:41:46 <Tofe> I would suspect our hwcomposer timeout -- something I didn't try so far was to deactivate that timeout completely
Feb 26 20:42:25 <saidinesh5> i see..
Feb 26 20:43:09 <Tofe> and that kind of logic is specific to LuneOS, so a bug there wouldn't necessarily be also in Sailfish
Feb 26 20:43:31 <saidinesh5> yeah
Feb 26 20:43:46 <saidinesh5> iirc sailfish uses something called mce to handle enabling/disabling the touchscreen
Feb 26 20:44:15 <Tofe> yes, it's provided by Mer, we don't use that one
Feb 26 20:44:22 <saidinesh5> Ah
Feb 26 20:45:05 <Tofe> we have mecanisms with powerd, sleepd, registered on LS2 bus, things like that
Feb 26 20:45:21 <Tofe> and the timing mecanism is in LunaSysMgr
Feb 26 20:45:32 <saidinesh5> Oh.. I see
Feb 26 20:45:56 <Tofe> Which has a nice State Machine to handle the different switches
Feb 26 20:46:22 <saidinesh5> Ahh.. interesting
Feb 26 20:46:51 <saidinesh5> this bug could just be a Qt level bug though right? the two touchscreen events one..
Feb 26 20:47:05 <Tofe> but I'll digg this deactivation of hwcomposer lead tomorrow, it's too late to begin that tonight :)
Feb 26 20:47:12 <saidinesh5> Aye
Feb 26 20:47:54 <Tofe> It could be; I would have to see exactly what event is sent by the kernel
Feb 26 20:48:36 <saidinesh5> yep.. like a simple Qt program inplace of the compositor.. should give you the idea
Feb 26 20:48:39 <Tofe> we have activated the evbug module, so it outputs the events it sees, which is a good thing here
Feb 26 20:48:53 <saidinesh5> Ahh
Feb 26 20:48:59 <saidinesh5> is it a kernel module?
Feb 26 20:49:34 <Tofe> more like a functionality
Feb 26 20:49:40 <saidinesh5> I see
Feb 26 20:50:05 <saidinesh5> the only time i had to deal with touchscreen issues was adding double tap support to my Mi 3 on sailfish os
Feb 26 20:50:07 <Tofe> it's really just a simple file in the kernel which adds hooks and printf code
Feb 26 20:50:13 <saidinesh5> Ahh
Feb 26 20:50:15 <saidinesh5> in sysfs?
Feb 26 20:50:20 <Tofe> dmesg
Feb 26 20:50:23 <saidinesh5> Ahh
Feb 26 20:51:35 <Tofe> Still, I'll have a look at sailfish's port for Mi 5, in case they added a commit of two
Feb 26 20:51:44 <saidinesh5> yeah
Feb 26 20:54:56 <saidinesh5> how's the integration/unification/not sure what to call it.. with webosse coming along?
Feb 26 20:55:10 <Tofe> lol, look at that last commit: https://github.com/zhxt/android_kernel_xiaomi_msm8996/commit/d4d941b1791e152f825f96589c8b9b3ac87020dc
Feb 26 20:56:01 <saidinesh5> Ah interesting
Feb 26 20:56:09 <saidinesh5> This could be something right?
Feb 26 20:56:34 <Tofe> could be; it's still a long shot, but who knows, it's in the right area
Feb 26 20:57:52 <saidinesh5> Yep
Feb 26 20:58:06 <saidinesh5> Didn't know the A1 had synaptics touchscreen
Feb 26 21:00:56 <Tofe> mmh it's already applied on my Mi 5 kernel -- my source code seems to be much more recent. But for me it's "rosy", whereas for sailfish it is "gemini", so a little bit different devices it seems
Feb 26 21:02:11 <Tofe> Anyway, time to go -- see you !
Feb 26 21:02:22 <saidinesh5> Good night?
Feb 26 21:02:36 <saidinesh5> I'm at 6890 packages
Feb 26 21:02:43 <saidinesh5> No question mark there btw. Lol
**** ENDING LOGGING AT Wed Feb 27 02:59:56 2019